THE SecurityWeek piece, written by Steve Durbin, argues that AI should augment rather than replace human security expertise, warning that autonomous cyber defence remains risky if treated as a guarantee of safety. It emphasises the danger of a closed loop where data quality, provenance, and model drift undermine decisions, creating a single point of systemic failure if humans are removed from oversight.
The article calls for transparency and governance, insisting we must know where AI is active, what data it uses, what decisions it can make, and when a human should intervene. It notes that true resilience depends on human intervention, with AI best used to surface weak signals and aid incident response, not to “put the pieces back together” after a major attack.
According to The United Nation’s Scientific Advisory Board, keeping pace with frontier AI capabilities will be critical, but pace must be balanced with governance and human judgement to avoid ceding control.